Monocultures and automobiles
Many have expressed concern about the use of networking technology in automobiles. Radio frequency identification chips (aka RFID chips) are common in the keychain "fobs" millions around the world use to open their cars. They're easy to use, and I've practically forgotten what it's like to have to put a key in my door to open it, but I'm not surprised that graduate students at Johns Hopkins University hacked into a Ford Escape using its a radio-frequency car key.
| John Carroll has shared his opinion on ZDNet since the last millennium. Since May, he's been a Microsoft employee. |
What piqued my interest, though, was the growing use of Bluetooth in high-end cars, as a recent article in Forbes ("Grand Theft Microchip," May 9, 2005) noted:
...Bluetooth networks are becoming a standard way to connect the car's stereo with the driver's phone and MP3 players. Half of Acura's fleet lets you wirelessly transfer a call between phone and car with a button on the steering wheel.
This is clearly cool stuff, and sure sign that convergence isn't just a buzz word bandied about at technology shows and marketing junkets. On the other hand, the more technology coalesces around a single standard (presuming more gets networked besides the radio), the more monoculture risks of the sort identified by Bruce Schneier come to the fore. Bluetooth viruses aren't just of academic interest. The "Cabir" test virus proved that such things are very possible, and other researchers have identified more vulnerabilities besides those used in Cabir.
Cost and benefit analysis should apply here, however. Diversification is the solution to monoculture risks. However, few would advocate diversifying beyond HTTP to encompass 10 or 20 networking protocols for Internet-based data interchange. We would be "more secure" in that those not using HTTP wouldn't be exposed to risks associated with HTTP networks (such as bugs in Web browsers that use HTTP as their networking protocol), but the costs of such a draconian solution, in terms of a less functional Web, less compatibility, and more development and maintenance effort would surely outweigh the risk.
The same applies to the creation of software and protocol "monocultures" in cars. Balkanized technology islands typifying early efforts at computing integration into automobiles pose less risks, but offer fewer integration opportunities. The development of monocultures, in other words, is critical if automobiles are going to be a part of the networked world alongside other "devices" in our lives.
So, when I read scare articles about the impending doom awaiting us from hijacked Hummers, I take it with a grain of salt. Yes, if I stuck with horse and buggies, there'd be less risk of collision injury. I'll still drive my car to work tomorrow.